Fliermodel ESC 22S 500A 25kW - Photos and Test Results

mkp007

100 mW
Joined
Oct 28, 2016
Messages
48
Location
San Diego, CA
'
I bought this new ESC from Fliermodel http://www.fliermodel.com/en/prc-show.asp?id=694

Mass = 1.52kg (as shown in photos below)
L x W x H: 7.8" x 4.3" x 2.0"

Many of you will likely want to reply that this thing is crap and will not withstand much power. And that might be so. However, please refrain from replying if you do not bring value to the discussion. There will be comparisons made with the testing of the 420A ESC https://endless-sphere.com/forums/viewtopic.php?f=2&t=44897.
I would love to get the really smart people with EE knowledge to chime in with what this ESC could be capable of doing given the info we have on this unit and past units.

I contacted the reps at Fliermodel and have been asking them for test data. One thing they have said is this can only do 25kW max. So 90V and 500A is not going to happen. I'm hoping for 80V and 250A for an input power of 20kW. Here is the test data they provided me using a 1.4m prop (still trying to confirm the details). As you can see, their battery is unable to support the amps. I hope this is why they did not push it further as that was about where the 420A ESC died.
Volts, Amps, Power (W), Thrust (kg)
86, 36, 3096, 21
82, 59, 4838, 30
78, 90, 7020, 38

Since the rep at Fliermodel strongly discouraged anything over 22S, I asked if it is because of the Mosfets and she said yes. So, the MOSFETs are likely 100V n-type, maybe like this one http://www.futurlec.com/Datasheet/Transistor/IRF540.pdf

In the image of the MCU board you see "48mos", does that mean 48 Mosfets? The side image does seem to show about that many (8 x 6). Notice how the IC labels are all scratched off. A total of 20, 100V 330uf capacitors.

Test results will be posted once I get the prop, motor and battery pack.

30905363935_b92747fff3.jpg


30154143624_fc81ef4822_z.jpg


30785904965_cbd20fcae6_b.jpg


30749271386_8f42f44cb2_c.jpg


30150221803_7fef1daab7_h.jpg
 
'
Since we have limited test data on these "high powered" ESC, I want compare everything we can with the Alien 24S 420A ESC and why it failed at 80V and 100A.

There is so much to consider:
1) Test voltage and current
2) Fet type and quantity
3) Capacitor type and quantity
4) Cooling fans, heat sinks
5) Motor kv, poles and rpm
6) ESC settings such as Timing and PWM freq
7) Size and weight
8 ) anything else I'm missing?

This picture shows the size of the Alien ESC. Not sure of the weight but my guess is about 620g (the FM-22S-500A ESC is 2x that). Bruno at Alien has made many changes to these ESC's over the years. here is a newer model http://alienpowersystem.com/shop/esc/alien-420a-4-24s-48-mosfet-esc-hv/
20121211_203838.jpg


Since it claimed to be 24S compatible, the fets would need to be higher rated than the 100V fets. If you review the description Alien's site he mentions 100V caps so not sure. There was plenty of discussion in the thread about using a lower voltage in the test to prevent early avalanche as the fets get hot. https://endless-sphere.com/forums/viewtopic.php?p=700565#p700679. I think this message was received as the test was limited to 90V (but difficult to verify by the test data).

Here are the test results:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=44897&start=600#p825135

From the video, 1st run to 5520rpm. The current reading is 28A and the output power is registering 4.1kW. Clearly something is wrong as this would indicate a voltage of about 150V. The motor is a JM1 and the kv on it is unknown so can not confirm the voltage that way. http://www.jobymotors.com/public/views/pages/jm1.php

30860220521_4f378ce878_b.jpg


2nd test and it appears to be bottoming out and then it fails. 6717RPM and 100A. Output voltage is 7.95kW which would indicate a battery voltage around 80V which seems right. Don't know if the torque reading is accurate as they have multiple set-ups depending on the motor mount being used. It's clear from the two videos that they don't have everything configured correctly in their software program.
30646235850_d091a315f3_b.jpg



This picture shows four cooling fans and two aluminum heat sinks. It appears the fet near the back end overheated and shorted out. The prop was pushing air into to. If the fans were pushing air in as well this would be a problem. They said they wired the fans to a separate power supply and could have easily made it so the fans were pushing air in. This could cause a stagnation zone in the region in question.
g+skrink+split+opened.jpg

pics and video by liveforphysics

So it is clear that one or more fets failed. Why did they fail? Heat? ok, probably. It would have been nice to put some T/C on those edge fets. If we look at the RPM and number of poles of the motor we can calculate how much switching those fets have to do. If we assume 24 poles and 6000rpm. That would be 12 pole pairs at 100rev/s so 1,200hz switching speed. Given a RDS, can we back out the amount of energy that needs to be dissipated per fet? But first we need to know how many fets and how many are in parallel, correct?

PMW needs to be 10x the switching speed so about 12khz would do. The following source says this: "Generally, raising the switching frequency increases PWM losses, though lowering the switching frequency limits the system’s bandwidth and can raise the ripple current pulses to the points where they become destructive or shut down the BLDC motor driver". hmmm, interesting. What was the ESC set to?
ref.
https://www.monolithicpower.com/Por.../appnotes/Brushless DC Motor Fundamentals.pdf

Other useful resources:
Mosfet Tutorial: http://www.microsemi.com/document-portal/doc_view/14692-mosfet-tutorial
Calculating Motor Driver Power Dissipation: http://www.ti.com/lit/an/slva504/slva504.pdf
Power dissipation in H-Bridge Motor Driver http://cache.nxp.com/files/analog/doc/app_note/AN5194.pdf
 
Looks like a significantly improved design!

Excited to see how it tests.
 
Its not clear, but i guess this is referin to the Alien ESC ?....
This picture shows four cooling fans and two aluminum heat sinks. It appears the fet near the back end overheated and shorted out. The prop was pushing air into to. If the fans were pushing air in as well this would be a problem. They said they wired the fans to a separate power supply and could have easily made it so the fans were pushing air in. This could cause a stagnation zone in the region in question.

But any progress / update on the Fliermodel 500a ESC ?
 
It failed from a control glitch causing shoot through IMHO.

Very common in my RC ESC experience to see them lose rotor sync for just a second or so and go from room temp to blasting plasma.
 
Hillhater said:
But any progress / update on the Fliermodel 500a ESC ?

Not yet. It will be another month or so.

It failed from a control glitch causing shoot through IMHO.

It would be nice to know for sure. Was heating and high voltage a factor? It seemed to be at steay state and then it just popped.

Here is a great post for those who want to know about the output waveforms of these ESCs http://www.dronevibes.com/forums/threads/product-review-jeti-hicopter-30a-opto-esc.20466/
 
mkp007 said:
'
Since we have limited test data on these "high powered" ESC, I want compare everything we can with the Alien 24S 420A ESC and why it failed at 80V and 100A.

There is so much to consider:
1) Test voltage and current
2) Fet type and quantity
3) Capacitor type and quantity
4) Cooling fans, heat sinks
5) Motor kv, poles and rpm
6) ESC settings such as Timing and PWM freq
7) Size and weight
8 ) anything else I'm missing?

This picture shows the size of the Alien ESC. Not sure of the weight but my guess is about 620g (the FM-22S-500A ESC is 2x that). Bruno at Alien has made many changes to these ESC's over the years. here is a newer model http://alienpowersystem.com/shop/esc/alien-420a-4-24s-48-mosfet-esc-hv/
20121211_203838.jpg


Since it claimed to be 24S compatible, the fets would need to be higher rated than the 100V fets. If you review the description Alien's site he mentions 100V caps so not sure. There was plenty of discussion in the thread about using a lower voltage in the test to prevent early avalanche as the fets get hot. https://endless-sphere.com/forums/viewtopic.php?p=700565#p700679. I think this message was received as the test was limited to 90V (but difficult to verify by the test data).

Here are the test results:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=44897&start=600#p825135

From the video, 1st run to 5520rpm. The current reading is 28A and the output power is registering 4.1kW. Clearly something is wrong as this would indicate a voltage of about 150V. The motor is a JM1 and the kv on it is unknown so can not confirm the voltage that way. http://www.jobymotors.com/public/views/pages/jm1.php

30860220521_4f378ce878_b.jpg


2nd test and it appears to be bottoming out and then it fails. 6717RPM and 100A. Output voltage is 7.95kW which would indicate a battery voltage around 80V which seems right. Don't know if the torque reading is accurate as they have multiple set-ups depending on the motor mount being used. It's clear from the two videos that they don't have everything configured correctly in their software program.
30646235850_d091a315f3_b.jpg



This picture shows four cooling fans and two aluminum heat sinks. It appears the fet near the back end overheated and shorted out. The prop was pushing air into to. If the fans were pushing air in as well this would be a problem. They said they wired the fans to a separate power supply and could have easily made it so the fans were pushing air in. This could cause a stagnation zone in the region in question.
g+skrink+split+opened.jpg

pics and video by liveforphysics

So it is clear that one or more fets failed. Why did they fail? Heat? ok, probably. It would have been nice to put some T/C on those edge fets. If we look at the RPM and number of poles of the motor we can calculate how much switching those fets have to do. If we assume 24 poles and 6000rpm. That would be 12 pole pairs at 100rev/s so 1,200hz switching speed. Given a RDS, can we back out the amount of energy that needs to be dissipated per fet? But first we need to know how many fets and how many are in parallel, correct?

PMW needs to be 10x the switching speed so about 12khz would do. The following source says this: "Generally, raising the switching frequency increases PWM losses, though lowering the switching frequency limits the system’s bandwidth and can raise the ripple current pulses to the points where they become destructive or shut down the BLDC motor driver". hmmm, interesting. What was the ESC set to?
ref.
https://www.monolithicpower.com/Por.../appnotes/Brushless DC Motor Fundamentals.pdf

Other useful resources:
Mosfet Tutorial: http://www.microsemi.com/document-portal/doc_view/14692-mosfet-tutorial
Calculating Motor Driver Power Dissipation: http://www.ti.com/lit/an/slva504/slva504.pdf
Power dissipation in H-Bridge Motor Driver http://cache.nxp.com/files/analog/doc/app_note/AN5194.pdf


So how did everything turned out at the end?
 
Back
Top